Implement context-length dependent KV-cache and Compute Buffer aware …#335
Merged
Nexesenex merged 1 commit intoJul 4, 2025
Conversation
…layer distribution for heterogeneous multi-GPU inference. Solves the problem of attemtping to run setups with different VRAM (e.g. 24GB cards with 6GB cards); previously layers were assigned without accounting for compute buffer, causing failure when one or more smaller GPUs could not hold the compute buffer. - Add requested_n_ctx parameter to llama_model_params - Implement 3-pass allocation algorithm accounting for compute buffers - Add device exclusion for insufficient memory (GPUs too small to allocate 1 layer + KV_cache + compute buffer excluded) - Add layer redistribution to make equitable use of included GPUs (may not be truly optimal)
Nexesenex
merged commit ccb3e6c
into
Nexesenex:lcpp_pr_kv_aware_layer_distrib
46 of 47 checks passed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
…layer distribution for heterogeneous multi-GPU inference. Solves the problem of attemtping to run setups with different VRAM (e.g. 24GB cards with 6GB cards); previously layers were assigned without accounting for compute buffer, causing failure when one or more smaller GPUs could not hold the compute buffer.